A track image recognition post-processing method based on curve fitting
A curve fitting and image recognition technology, applied in the field of rail transit, can solve the problems of uneven track edges, large differences, misidentification, etc., and achieve the effect of precise positioning
Active Publication Date: 2019-05-17
浙江众合科技股份有限公司
View PDF10 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the recognition result is a combination of pixels, and the obtained track edge is not smooth, which is quite different from the actual one.
And sometimes other objects beside the track are identified as pixels of the track, and misidentification occurs
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0031] The technical problem to be solved by the present invention is to provide a track image post-processing method based on curve fitting to correct the image result identified by the deep learning method.
[0032] The specific process of the whole method is:
[0033] S1, input the original image img_input, the original image img_input is the image obtained after track recognition through the deep learning method, and binarize the original image to obtain the binary image img_binary;
[0034] S2, find the maximum connected domain in the binary image img_binary, set the points of other non-maximum connected domains as background points, and output the image img_maxDomain;
[0035] S3, find the left and right track lines in the image img_maxDomain;
[0036] S4. Limiting filtering is performed on the left and right orbital lines, and the points that deviate far away from each orbital line are filtered out;
[0037] S5, draw the trajectory line according to the curve fitting ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a track image recognition post-processing method based on curve fitting. The method comprises the following steps of S1, inputting an original image img _ input, wherein the original image img _ input is an image obtained after track recognition is conducted through a deep learning method, carrying out the binaryzation on the original image, and obtaining a binary image img_ binaryzation; S2, finding out a maximum connected domain in the binary image img _ binary, setting points of other non-maximum connected domains as background points, and outputting an image img _maxDomain; S3, finding a left track line and a right track line in the image img _ maxDomain; S4, carrying out the amplitude limiting filtering on the left rail line and the right rail line, and filtering out the points far away from each rail line; and S5, drawing a track line according to the curve fitting function. According to the image post-processing method based on curve fitting, the recognition result can be corrected, and the image closer to the real track can be drawn.
Description
technical field [0001] The invention relates to rail transit technology, in particular to rail transit image processing technology. Background technique [0002] In the field of rail transit, based on the deep learning algorithm, the trained network model can be used to initially identify the track from the image. However, the recognition result is a combination of pixels, and the obtained track edge is not smooth, which is quite different from the actual one. And sometimes other objects beside the track are identified as the pixels of the track, and misidentification occurs. Contents of the invention [0003] The technical problem to be solved by the present invention is to provide a track image post-processing method based on curve fitting to correct the image result identified by the deep learning method and draw an image closer to the real track. [0004] In order to solve the above-mentioned technical problems, the present invention adopts the following technical so...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/34G06K9/00
Inventor 程艳丽饶双宜
Owner 浙江众合科技股份有限公司